Preston Curve

library(tidyverse)
library(ggrepel)
library(plotly)
library(styler)

Import CSV as tibble

life_quality <- read_csv("life_quality.csv")

Make a bubble chart with ggplot2

Q2 <- ggplot(
  data = life_quality,
  aes(
    x = gdp_per_capita,
    y = life_expectancy,
    label = country_name
  )
) +
  geom_point(aes(
    size = pop,
    colour = continent
  ),
  # make the points semi-transparent 
  alpha = 0.5
  ) +
  # add titles 
  labs(
    x = "GDP per capita (US$, PPP 2015)",
    y = "Life expectancy (years)",
    title = "Wealth and Health by Country",
    caption = "Source : World Bank"
  ) +
  # adjust the title to the center 
  theme(plot.title = element_text(hjust = 0.6))

Q2

Change the x-coordinates to a logarithmic scale

Q3 <- Q2 + 
  scale_x_continuous(trans='log10',
                 breaks = c(1000,10000, 100000),
                 # add labels 
                 labels = c("$1,000", "$10,000", "$100,000"),
                 # add minot ticks 
                 minor_breaks = c(seq(0,1000, 100),
                                  seq(1000, 10000, 1000), 
                                  seq(10000, 100000, 10000)))

Q3

Use scale_size_area() so that the areas of the bubbles in the legend represent populations of 1 million,10 million, 100 million and 1 billion. Change the numbers in the legends from ‘1e+06’, ‘1e+07’, ‘1e+08’,‘1e+09’ to the reader-friendlier strings ‘1 million’, ‘10 million’, ‘100 million’ and ‘1 billion’. Increase max_size so that the bubble areas are approximately the same as in the Gapminder figure.
```
Q4 <- Q3 +
 scale_size(
 range = c(1, 1000)
 ) +
 scale_size_continuous(
 name = "Population",
 range = c(1,20),
 breaks = 10^(6:9),
 labels = c(
 "1 million", 
 "10 million", 
 "100 million", 
 "1 billion"
 )
 )
Q4
```
Change the colour scale to the ColorBrewer palette ‘Set1’.
```
Q5 <- Q4 + 
 scale_colour_brewer(palette = "Set1")
Q5
```

Override the alpha aesthetic in the legend. Also increase the sizes of the circles in the colour legend so that the colours are easier to read.

Q6 <- Q5 +
  guides(colour = guide_legend(override.aes = list(alpha = 1, size = 3)))+
  # Change the legend title 
  scale_colour_discrete("Continents")
Q6

Fit a single LOESS curve to all data points.

Q7 <- Q6 +
  geom_smooth(
    method = "loess",
    aes(weight = pop),
    color = "darkgrey"
)   
Q7

Using ggrepel::geom_text_repel() to add country names as labels

long_lived <- life_quality |>
  slice_max(life_expectancy, n = 5)

short_lived <- life_quality |>
  slice_min(life_expectancy, n = 5)

wealthiest <- life_quality |>
  slice_max(gdp_per_capita, n = 5)

poorest <- life_quality |>
  slice_min(gdp_per_capita, n = 5)

fertile <- life_quality |>
  slice_max(pop, n = 5)

# combine all and remove overlapped countries 
life_quality_selected <- 
  distinct(
    rbind(
      long_lived,
      short_lived,
      wealthiest,
      poorest,
      fertile
      )
    )

Q8 <- Q7 + 
  geom_text_repel(
    data = life_quality_selected, 
    aes(
      label = country_name, 
      color = continent
      ), 
    # set the point size 
    size = 3,
    # reduce the overlap between labels 
    force = 100
    )
Q8

make more adjustments if you think they improve the quality of the plot.
```
Q9 <- ggplotly(Q7)
Q9
```
Here, we plotted an interactive plot using the function ggplotly(). By hovering the pointer over a country in the graph, an individual will be able to see all the information of a country. However, the labels in the hovertext are untidy and thus we can use the following code in the bottom:
```
data <- distinct(life_quality)
data <- data[order(data$continent, data$country_name),]
slope <- 1e-06
data$size <- sqrt(data$pop * slope)
colors <- c('#4AC6B7', '#1972A4', '#965F8A', '#FF7070', '#C61951')
fig <- plot_ly(data, 
 x = ~log10(gdp_per_capita), 
 y = ~life_expectancy, 
 color = ~continent, 
 size = ~size, 
 colors = colors,
 type = 'scatter',
 mode = 'markers',
 sizes = c(1, 50),
 marker = list(symbol = 'circle',
 sizemode = 'diameter',
 line = list(width = 2, 
 color = '#FFFFFF')),
 text = ~paste('Country:', country_name,
 ' Life Expectancy:', life_expectancy,
 ' GDP:', gdp_per_capita,
 ' Pop.:', pop,
 ' Continent:', continent)) |>
 layout(title = 'Life Expectancy v. Per Capita GDP',
 xaxis = list(title = 'GDP per capita',
 type = "log",
 gridcolor = 'rgb(255, 255, 255)',
 zerolinewidth = 1,
 ticklen = 5,
 gridwidth = 2),
 yaxis = list(title = 'Life Expectancy (years)',
 gridcolor = 'rgb(255, 255, 255)',
 zerolinewidth = 1,
 ticklen = 5,
 gridwith = 2),
 paper_bgcolor = 'rgb(243, 243, 243)',
 plot_bgcolor = 'rgb(243, 243, 243)',
 legend = list(itemsizing = 'constant',
 title = list(text = ' Continents '))) |>
 animation_opts(2000,
 redraw = FALSE) |>
 add_trace(y = ~fitted(loess(life_expectancy ~ log10(gdp_per_capita))),
 line = list(color = colors),
 showlegend = F)
fig
```
The code is written with reference to the website Plotly ¹. Here, there 5 lowess lines, each line representing one continent. While the points representing the countries are smaller, they are more proportional to its population. To solve the issue of the points being too small, this plot enables users to zoom into a specific area of the plot that they want to investigate.
Write a few sentences about the data. What does the plot reveal about the data?
The plot shows that there is an overall trend where as the Gross Domestic Product (GDP) per capita increases, the life expectancy of the country increases. From the graph, we also see that the countries part of the Africa continent tend to have lower life expectancy and lower GDP per capita while countries from the Europe continent tend to have higher life expectancy and higher GDP per capita.

https://plotly.com/python/bubble-charts/. Accessed 21st February 2022.↩︎

Preston Curve

Team A

2/16/2022